Rules to Tools: Executable Checks for LLM Agents in Scientific Computing

DOI10.48550/arXiv.2610.00313MaRDI QIDQ7367483arXivFDO

Authors Jingjie Ning, Ji Zeng, Guolin Ke, Guojiang Zhao, Chen Xu, Shanshan Zhong, Xiaochuan Li

Publication date 29 September 2026

Published in arXiv

Tldr We provide an executable tool to check scientific computing requirements against written programs; across two task sets, our tool group achieves 26/30 complete repairs versus the text-only group's 29/30.

Full work available at URL https://doi.org/10.48550/arXiv.2610.00313

arXiv classification

cs.AIcs.SE

This page was built for publication: Rules to Tools: Executable Checks for LLM Agents in Scientific Computing

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7367483)